The FcϵRIβ Immunoreceptor Tyrosine-based Activation Motif Exerts Inhibitory Control on MAPK and IκB Kinase Phosphorylation and Mast Cell Cytokine Production*

The high affinity IgE Fc receptor (FcϵRI) β chain functions as a signal amplifier and has been linked to atopy, asthma, and allergy. Herein, we report on a previously unrecognized negative regulatory role for the nonconventional β chain immunoreceptor tyrosine-based activation motif that contains three tyrosine residues (YX5YX3Y). Degranulation and leukotriene production was found to be impaired in cells expressing the mutated FcϵRIβ immunoreceptor tyrosine-based activation motifs FYY, YYF, FYF, and FFF. In contrast, cytokine synthesis and secretion were enhanced in the YFY and FFF mutants. FcϵRI phosphorylation and Lyn kinase co-immunoprecipitation was intact in the YFY mutant but was lost in the FYF and FFF mutants. The phosphorylation of Syk, LAT, phospholipase γ1/2, and Srchomology 2 domain-containing protein phosphatase 2 was intact, whereas the phosphorylation of SHIP-1 was significantly reduced in the YFY mutant cells. The FYF and FFF mutants were defective in phosphorylating all of these molecules. In contrast, the phosphorylation of ERK, p38 MAPK, IκB kinase β (IKKβ), and nuclear NFκB activity was enhanced in the YFY and FFF mutants. These findings show that the FcϵRIβ functions to both selectively amplify (degranulation and leukotriene secretion) and dampen (lymphokine) mast cell effector responses.

The high affinity IgE Fc receptor (Fc⑀RI) ␤ chain functions as a signal amplifier and has been linked to atopy, asthma, and allergy. Herein, we report on a previously unrecognized negative regulatory role for the nonconventional ␤ chain immunoreceptor tyrosine-based activation motif that contains three tyrosine residues (YX 5 YX 3 Y). Degranulation and leukotriene production was found to be impaired in cells expressing the mutated Fc⑀RI␤ immunoreceptor tyrosine-based activation motifs FYY, YYF, FYF, and FFF. In contrast, cytokine synthesis and secretion were enhanced in the YFY and FFF mutants. Fc⑀RI phosphorylation and Lyn kinase co-immunoprecipitation was intact in the YFY mutant but was lost in the FYF and FFF mutants. The phosphorylation of Syk, LAT, phospholipase ␥1/2, and Srchomology 2 domain-containing protein phosphatase 2 was intact, whereas the phosphorylation of SHIP-1 was significantly reduced in the YFY mutant cells. The FYF and FFF mutants were defective in phosphorylating all of these molecules. In contrast, the phosphorylation of ERK, p38 MAPK, IB kinase ␤ (IKK␤), and nuclear NFB activity was enhanced in the YFY and FFF mutants. These findings show that the Fc⑀RI␤ functions to both selectively amplify (degranulation and leukotriene secretion) and dampen (lymphokine) mast cell effector responses.
The Fc⑀RI 1 in mast cells and basophils is a tetrameric structure consisting of three distinct polypeptides including the IgE-binding ␣ chain, the tetraspanning ␤ chain, and the disulfide-linked ␥ homodimer (1). Aggregation of Fc⑀RI by the interaction of bound IgE with multivalent antigens induces the release of histamine, leukotrienes, and inflammatory cytokines, resulting in the recruitment and activation of circulating leukocytes leading to allergic inflammation (2). Polymorphisms in the Fc⑀RI␤ have been linked to atopy, asthma, and allergy (3)(4)(5). However, studies on the role of these polymorphisms in Fc⑀RI expression and function have not yielded an understanding of their effects on mast cell physiology (6,7). Nonetheless, it is clear that the Fc⑀RI␤ functions as a signal amplifier in mast cells and is important for augmenting the allergic reactions (8 -10). In humans, it has been demonstrated that the ␥ homodimer can act as an autonomous signaling molecule in various cell types, whereas the ␤ subunit amplifies early activation signals by 5-7-fold through Fc⑀RI (reviewed in Ref. 1). Moreover, the ␤ chain also enhances cell-surface expression of Fc⑀RI (11), providing increased sensitivity in an allergic response.
The ␤and ␥-chains contain ITAMs, a conserved feature of many antigen receptors that imparts signaling competence. The ITAM consensus sequence is (D/E)XXYXXLX 7-11 -YXXL(L/ I), where the tyrosine residues are phosphoacceptor sites for the action of receptor-associated protein tyrosine kinases (reviewed in Ref. 1). Phospho-ITAMs provide a docking site for cytoplasmic proteins that contain the Src-homology 2 (SH2) domain. Some of these function to couple receptors to macromolecular signaling complexes anchored by adapter proteins (12), which serve to coordinate and amplify intracellular signals. The structure of the ␤ and ␥ chain ITAMs differ. In particular, the ␤ chain ITAM shows a notable departure from the consensus ITAM sequence with the presence of a third tyrosine (the middle tyrosine) between the two canonical tyrosine residues and a shorter spacer region between the canonical tyrosines. The Src family protein tyrosine kinase Lyn weakly binds to ␤ chain ITAM in resting cells and is further recruited after the receptor aggregation (13)(14)(15). Lyn is required for the initial step of activation by phosphorylating Fc⑀RI specifically on the ␤ and ␥ ITAMs (16,17). Although the requirements for this initial event are not fully understood, it appears that the protein-protein interaction of Lyn with Fc⑀RI and an appropriate lipid microenvironment, where this interaction may be enhanced, are probably prerequisites (18 -20). Recently, we described that another Src family protein tyrosine kinase Fyn was also activated by and co-immunoprecipitated with the Fc⑀RI (21). This kinase initiates a complementary signaling pathway that is anchored by the adapter molecule Gab2 (Grb2-associated binder protein 2) whose phosphorylation is critical for activation of phosphatidylinositol 3-OH ki-nase (PI3K) and for membrane targeting of the SH2 domaincontaining protein phosphatase 2 (SHP-2) (22). However, because the requirements for Fc⑀RI-dependent Fyn activation are yet unclear, in this study, we focused on testing the role of the Fc⑀RI␤ ITAM tyrosines in Lyn-dependent activation of mast cell responses. This focus was based on the recent finding that Lyn kinase is a negative regulator of mast cell responsiveness and the allergic response (23), thus raising the possibility that its interaction with Fc⑀RI␤ could exert negative regulatory influences.
Because bone marrow-derived cultured mast cells (BMMC) from ␤ chain-deficient mice lack cell surface expression of Fc⑀RI despite normal expression of Fc⑀RI ␣ and ␥ chain mRNA, we chose this system to retrovirally transduce mutant ␤ chains that reconstituted cell surface Fc⑀RI. Most mutant Fc⑀RI␤expressing cells showed impaired degranulation and leukotriene secretion when compared with wild type Fc⑀RI␤-expressing cells. The YFY mutant was an exception, as it showed no inhibition of these responses. Strikingly, cytokine synthesis and secretion was increased in YFY mutant cells and was greatly enhanced in the FFF mutant. The increased cytokine production was associated with increased activation of the gene transcription regulators, MAPKs, and NFB. Whereas a positive role for the Fc⑀RI ␤ chain in allergic responses has been rigorously demonstrated (reviewed in Ref. 1), our findings revealed a previously unrecognized negative regulatory role for the Fc⑀RI␤ ITAM in cytokine production.

EXPERIMENTAL PROCEDURES
Antibodies and Reagents-Anti-rat ␤ monoclonal Ab (clone JRK), which also recognizes mouse ␤ chain (24), was used in these studies. Mouse monoclonal Abs to IgE (anti-TNP or anti-DNP) were also used (25). IgE was labeled with FITC (Sigma). Mouse monoclonal Abs to SHP-2 were purchased from BD Biosciences. Recombinant murine stem cell factor and interleukin-3 (IL-3) were kindly gifted by Kirin Brewery Co. (Tokyo, Japan) or were purchased from PeproTech (Rocky Hill, NJ). Polyclonal Abs to PI3K and SHIP and a biotinylated monoclonal antibody to phosphotyrosine (clone 4G10) were from Upstate Biotechnologies (Lake Placid, NY). Rabbit polyclonal Abs to Lyn and PLC␥1 and monoclonal antibody to PLC␥2 were from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit polyclonal and mouse monoclonal Abs to Syk were kindly provided by U. Blank (Institute Pasteur, Paris, France) and P. Draber (Institute of Molecular Genetics, Prague, Czechoslovakia), respectively. Polyclonal Ab to LAT was generously gifted by L. E. Samelson (National Institutes of Health, Bethesda, MD). All of the phosphospecific Abs were from Cell Signaling Technology (Beverly, MA) as well as antibodies to IKK␣/␤ and IB. Horseradish peroxidaseconjugated extravidin and monoclonal anti-rabbit immunoglobulin (clone RG16) were from Sigma. Horseradish peroxidase-conjugated antibodies to mouse and rabbit immunoglobulin were from Amersham Biosciences. Detergents (Nonidet P-40 and octyl glucoside) were from Pierce. Phosphatase and proteinase inhibitors were from Sigma. NaF was from Fluka (Buchs, Switzerland).
Mice-The ␤ chain-deficient (␤ Ϫ/Ϫ ) mice on the chimeric BDF 1 and 129 background have been described previously (10) and were bred in the animal facility under specific pathogen-free conditions. BMMCs from ␤ ϩ/ϩ mice from the same background as ␤ Ϫ/Ϫ mice were used as controls. For some experiments, BMMCs from these mice were kindly provided by Drs. Toshiaki and Yuko Kawakami (La Jolla Institute of Allergy and Immunology).
Cell Culture and Retroviral Gene Transduction-Bone marrow was harvested from mice femurs, and BMMCs were cultured as described previously (26). After 4-weeks of culture, cells were stained for cell surface expression of c-kit and/or Fc⑀RI and used for experiments when Ͼ95% cells showed mast cell markers. For retroviral gene transduction, bone marrow cells were grown for 8 -10 days in the presence of IL-3 before co-culturing with a mixture of the ecotropic retrovirus-packaging cell lines, Phoenix and BOSC. After 2 days of co-culture, BMMC were recovered by centrifugation and antibiotic selection (puromycin) was initiated as described previously (27). After 2 weeks of selection, cells were analyzed for Fc⑀RI expression. Cultures were used when Ͼ95% cells expressed Fc⑀RI.
Site-specific Mutagenesis and Retroviral Constructs-The PCR-based mutagenesis described by M. P. Weiner et al. (28) was employed to create single nucleotide-mutagenized constructs. The 800-bp EcoRI fragment containing the mouse ␤ chain coding sequence was isolated and subcloned into the EcoRI site of pBluescript (Stratagene, La Jolla, CA). Two sets of PCR primers, including the mutation site, were designed in the opposite direction of the mutation site. The sequences of sense and antisense oligonucleotide primers for each mutagenesis were as follows: wild type, 5Ј-G TAT TCA CCA ATT TAC AGT GAG TT-3Ј  and 5Ј-AC ATT TAA TTC TTC ATA AAG ACG AT-3Ј;  PCR was carried out by using 2.5 units of Thermococcus kodakaraensis polymerase (Toyobo, Tokyo, Japan) in a 100-l reaction mixture containing 10 ng of pBluescript as template, 50 pM of each phosphorylated primer, 2 mM MgCl 2 , 0.2 mM dNTPs, 120 mM Tris-HCl (pH 8.8), 10 mM KCl, 6 mM (NH 4 ) 2 SO 4 , 0.1% Triton X-100, and 10 g/ml BSA). Amplified samples were denatured at 94°C for 4 min before 25 cycles of amplification (94°C for 1 min, 55°C for 1 min, and 72°C for 4 min) followed by a final extension at 72°C for 10 min. The PCR products were treated with DpnI (New England Biolabs, Beverly, MA) to digest the template DNA followed by self-ligation. The integrity of all of the mutants was confirmed by sequence analysis using capillary sequencer (ABI prism 310 Genetic Analyzer, ABI, Foster City, CA). The mutant ␤ chain cDNAs were subcloned into the EcoRI site of the Moloney murine leukemia virus-based retroviral vector, pMX-puro, and the virus was produced (29). Empty vector (pMX-puro) was used as a negative control to generate the virus not expressing the Fc⑀RI␤.
Flow Cytometry-The cells (3 ϫ 10 5 ) were incubated with 5 g/ml FITC-labeled IgE on ice for 30 min in 200 l of phosphate-buffered saline (PBS) with 0.1% BSA and 0.05% sodium azide. The cell surface fluorescence intensity was measured using a flow cytometer (FACScan R , BD Biosciences). The number of dead cells was determined by propidium iodide staining.
␤-Hexosaminidase Release Assay-Degranulation via Fc⑀RI stimulation was determined by ␤-hexosaminidase release. Cells (5 ϫ 10 5 cells/ ml) were incubated with TNP or DNP-specific IgE (2 g/ml) in Mg 2ϩ and Ca 2ϩ -free Tyrode's buffer (10 mM Hepes buffer, pH 7.4, 130 mM NaCl, 5 mM KCl, 5.6 mM glucose, and 0.1% BSA) for 30 min at 4°C. The sensitized cells were washed and resuspended at a concentration of 5 ϫ 10 5 cells/ml in Tyrode's buffer containing 1 mM CaCl 2 and 0.6 mM MgCl 2 . TNP-ovalbumin (1 ng/ml) or DNP-BSA (10 ng/ml) was used for maximal stimulation or at indicated concentrations. Phorbol 12-myristate 13-acetate (10 ng/ml) and ionomycin (100 ng/ml) were employed for pharmacological stimulation as positive control. After a 30-min incubation at 37°C, the reaction was terminated by centrifugation at 4°C. The total hexosaminidase concentration was obtained by cell lysis in 1% Nonidet P-40. Aliquots of the supernatants and total cell lysates were incubated with 1.3 mg/ml p-nitrophenyl-N-acetyl-␤-D-glucopyranoside (Sigma) in 0.1 M sodium citrate buffer (pH 4.5) for 90 min at 37°C. The reaction was terminated by the addition of 0.2 M glycine buffer (pH 10.7). The release of the product 4-p-nitrophenol was monitored by optical absorbance at 405 nm. Percentage of ␤-hexosaminidase release was calculated as follows: (supernatant optical density value of the stimulated cells Ϫ supernatant optical density value of the unstimulated cells) ϫ 100/(the total cell lysate optical density value Ϫ supernatant optical density value of the unstimulated cells). Percent spontaneous release was usually below 5%.
Cytokine, LTC 4 , NFB Activity, and Real-time PCR-Cell culture supernatants after stimulation were analyzed for IL-6, IL-13, TNF␣, and LTC 4 by specific ELISA kits according to the manufacturers' instruction (Genzyme Techne, Minneapolis, MN or BIOSOURCE International for IL-6, IL-13, and TNF␣, Cayman Chemical, Ann Arbor, MI for LTC 4 ). Cells (2 ϫ 10 5 cells/ml) sensitized with IgE (2 g/ml) were stimulated with antigen for 3 h at 37°C, and supernatants collected and analyzed. NFB binding activity was determined by measuring p65 binding to a NFB consensus binding site oligonucleotide in an ELISAbased format according to the manufacturers' instructions (TransAM TM NFB family, Active Motif, Carlsbad, CA). Nuclear lysates (prepared per manufacturers' instruction) of stimulated and nonstimulated (as above) cells expressing the YYY, YFY, FYF, and FFF ITAMs were incubated in ELISA wells containing the NFB consensus binding site oligonucleotide. Binding was measured by incubation with a primary antibody to p65 followed by a secondary horseradish peroxidase-conjugated antibody that recognized the primary antibody. Following a final step of multiple washes, a colorimetric reaction was performed. The positive control of stimulated Jurkat T cells and negative lysis buffer controls were also included. Absorbance was read at 450 nm.
For analysis of mRNA levels, RNA was isolated using a Qiagen RNeasy kit (Valencia, CA) and reverse-transcribed using a first-strand cDNA synthesis kit (Invitrogen). Real-time PCR (TaqMan) was performed using the ABI PRISM7700 sequence detection system (Applied Biosystems, Foster City, CA). An analysis of IL-6, TNF␣, and 18 S rRNA levels was performed using commercially available primer/probe sets (Applied Biosystems). Levels of IL-6 and TNF␣ were determined by normalization of 18 S rRNA levels relative to the unstimulated transduced cell line, which was arbitrarily designated a value of 1.0.
Cell Lysates, Immunoprecipitation, Immunoblotting, and Phospho-ITAM Peptide Pull-down Assays-Cell aliquots of 3 ϫ 10 7 were stimulated in a total volume of 1 ml of Tyrode buffer and sampled at different times or stimulated with 100 ng/ml antigen. Stimulated cells were washed twice with ice-cold PBS and lysed in either SDS sample buffer or borate-buffered saline that contained 1% Nonidet P-40, 60 mM octyl glucoside, 2 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 2 g/ml leupeptin and pepstatin, 5 mM sodium pyrophosphate, 50 mM NaF, and 1 mM sodium orthovanadate for 30 min on ice. Lysates were centrifuged for 20 min (4°C) at 14,000 rpm. For immunoprecipitation, cell lysates were incubated for 3 h with antibodies prebound to either protein G-Sepharose (monoclonal) or protein A-Sepharose (polyclonal). Proteins were recovered by denaturation with an equal volume of 2ϫ Tris-glycine SDS sample buffer that contained 1% 2-mercaptoethanol and 1 mM orthovanadate and were resolved by 8, 10, or 12% SDS-PAGE (Invitrogen). Nonreducing (lacking 1% 2-mercaptoethanol) SDS-PAGE was used for resolving Fc⑀RI immunoprecipitates to avoid masking of the receptor subunits by immunoglobulin heavy and light chains in Western blots. Proteins were electrophoretically transferred to nitrocellulose membranes (from Invitrogen or Schleicher & Schuell). Nitrocellulose membranes were blocked with either 4% BSA or 5% dry milk (depending on the primary antibody) in Tris-buffered saline containing 0.1% Tween 20. They were then probed with the desired primary antibody and an appropriate secondary horseradish peroxidase-conjugated antibody and visualized by enhanced chemiluminescence.
To confirm and extend co-immunoprecipitation experiments for proteins interacting with Fc⑀RI␤ ITAM, pull-down assays were employed using biotinylated phospho-ITAM peptides (pY-pY-pY, pY-Y-pY, Y-pY-Y, YYY; AC Scientific, Duluth, GA), equivalent to YYY, YFY, FYF, and FFF, bound to streptavidin-Sepharose beads (Fluka Biochemika, Switzerland). For these experiments, peptide-containing beads were incubated with cell lysates prepared in radioimmune precipitation assay buffer (10 mM NaH 2 PO 4 , 50 mM NaCl, 50 mM NaF, 5 mM EDTA, 0.5% deoxycholic acid, 0.05% NaN 3 , 0.1% SDS, 1% Triton X-100, and 1 mM Na 3 VO 4 ) from IgE/Ag-activated or nonactivated BMMC for 3 h at 4°C. Bead pellets were recovered by microcentrifugation and washed 3ϫ with radioimmune precipitation assay, and bound proteins were recovered by extraction in SDS sample buffer. Recovered proteins were resolved by SDS-PAGE and transferred as above for immunoblotting. No differences in protein interactions with the Fc⑀RI␤ ITAM were noted when stimulated or nonstimulated cell lysates were used.
Calcium Measurements-Calcium was measured as described previously (26). Cells (2 ϫ 10 6 ) were dual-loaded with 16 M Fluo-4-AM and 16 M Fura Red (Molecular Probes, Eugene, OR) in RPMI 1640, 2% fetal calf serum media for 45 min at 37°C. Cells were incubated with IgE (1 g/10 6 cells) on ice for 1 h and brought to room temperature for 20 min. Cells were resuspended in Tyrode-BSA, and changes in dye fluorescence with time were determined by flow cytometry after stimulation with 30 ng/ml antigen at 37°C. Calcium mobilization was reported as the ratio of Fluo-4 to Fura Red fluorescence intensity over time.
Statistical Analysis-A nonparametric test (Mann-Whitney) was employed to determine the statistical significance among the experimental groups. p Ͻ 0.05 was considered significant.

Expression of Wild Type and Mutant Forms of Fc⑀RI␤ in BMMC from ␤ Chain-deficient Mice Reconstitutes Cell Surface
Expression of Fc⑀RI-To investigate the functional significance of canonical and noncanonical tyrosine of the ␤ chain ITAM in mast cell responses, we generated five mutants of the ␤ chain with the replacement of tyrosine to phenylalanine in which the YYY (N-terminal to C-terminal sequence) was changed to YYF, YFY, FYY, FYF, and FFF. Retroviral-mediated gene transfer into ␤ Ϫ/Ϫ mouse mast cells and cell surface expression analysis of Fc⑀RI revealed that cells expressing the YYF, YFY, FYY, FYF, and FFF were all Fc⑀RI-positive on their cell surface (Ͼ95%), similar to cells transduced with the wild type (YYY) ITAM virus, whereas cells transduced with the empty vector (pMX-puro) virus did not express Fc⑀RI, similar to ␤ Ϫ/Ϫ mast cells (Fig. 1A). The relative fluorescence intensities of Fc⑀RI for each transduced cell culture were almost identical. The cells were shown to express Fc⑀RI␤ protein by immunoblotting with monoclonal antibody to ␤ chain (Fig. 1B). The 31-kDa band, which corresponds to the ␤ chain, in YYF, YFY, FYY, FYF, FFF, and wild type (YYY) cells was easily detected, whereas no protein band was observed in cells transduced with the vectorcontrol virus (Fig. 1B). Moreover, the level of protein expression was similar to nontransduced BMMC. These results show that the mutated ␤ chain ITAM successfully reconstituted Fc⑀RI cell surface expression to the same extent as wild type ITAM, and expression levels were comparable among the transduced cells.
Mutation of the Fc⑀RI ␤ ITAM Canonical Tyrosines but Not the Noncanonical Tyrosine Partly Impairs Mast Cell Degranulation and LTC 4 Secretion-We investigated whether the expression of the different mutant ␤ chains affected the degranulation and LTC 4 secretion responses of the transduced mast cells upon Fc⑀RI stimulation. As shown in Fig. 2A, all of the cells expressing surface Fc⑀RI with wild type or mutant ␤ chains, but not pMX-puro-transduced cells, were able to degranulate in response to Fc⑀RI stimulation. Cells expressing the FYY, YYF, FYF, and FFF forms of the ␤ chain ITAM showed a 30 -50% reduction in mast cell degranulation and LTC 4 secretion (Fig. 2B) when compared with wild type ITAMtransduced cells (YYY) or wild type BMMC (data not shown). This is consistent with the previously described amplifying function of Fc⑀RI␤ on mast cell degranulation (9). In contrast, there was no significant difference in the degranulation or LTC 4 response of the YFY mutant when compared with wild type ITAM-transduced cells or BMMC, suggesting that the noncanonical tyrosine does not contribute to Fc⑀RI␤-mediated amplification.
Fc⑀RI-mediated Cytokine Production and Secretion Is Dramatically Enhanced in the Absence of Fc⑀RI ␤ ITAM Tyrosines-IgE-dependent activation of mast cells leads to de novo synthesis and secretion of cytokines (30). We investigated whether mutations of the ␤ chain ITAM would influence the cytokine secretion from mast cells. Because IL-6, IL-13, and TNF␣ are cytokines that are potently produced and secreted from mast cells, we measured their release by ELISA (Fig. 3, A-C). As expected, the engagement of the Fc⑀RI carrying the wild type ␤ chain ITAM induced a significant net release of IL-6, IL-13, and TNF␣ when compared with pMX-puro-transduced cells (where cytokines measured from 5 to 10% of response of wild type ␤ chain-reconstituted cells, see Fig. 3 legend). The YYF-, FYY-, and FYF-expressing cells either showed no significant change or a significant decrease in response when compared with the wild type ITAM-expressing cells (Fig.  3, A-C). Interestingly, cells expressing the mutant of the noncanonical tyrosine (YFY) showed a modestly elevated level of IL-6 and IL-13 secretion when compared with wild type cells, whereas TNF␣ was not significantly affected. This differed from the results for ␤-hexosaminidase release and LTC 4 secretion where the YFY mutant showed no significant difference when compared with the wild type ITAM, suggesting that the noncanonical tyrosine could contribute to the cytokine response. A striking enhancement in IL-6, IL-13, and TNF␣ secretion was observed in cells expressing the FFF mutant.
Cytokine secretion was enhanced by 2-4-fold compared with cells expressing the wild type ITAM (YYY). To further elucidate whether this negative regulatory role of the Fc⑀RI ␤ ITAM was at the level of protein secretion or mRNA expression, we examined the cytokine mRNA expression of YYY, YFY, FYF, and FFF ITAM-expressing cells by real-time PCR (TaqMan). The expression of mRNA for IL-6 and TNF␣ in YFY and FFFexpressing cells was increased post-Fc⑀RI stimulation when compared with the wild type (YYY) ITAM-expressing cells (Fig.  3D). Although the increase in TNF␣ mRNA was quite modest in YFY-expressing cells, it was significant and reproducible. In contrast, FYF-expressing cells showed reduced mRNA levels relative to the wild type ITAM. This finding suggested that the mutation of the noncanonical tyrosine in the Fc⑀RI␤ ITAM released a negative regulatory constraint on gene expression that is augmented by (but does not depend on) the canonical tyrosines since the FFF, but not the FYF, mutant demonstrated this phenotype.
Phosphorylation of Fc⑀RI ␤ Chain and ␥ Chains, Lyn Interaction with the ␤ Chain, and Phosphorylation of Receptorproximal Signaling Molecules Are Partly Dependent on an Intact Fc⑀RI ␤ Chain ITAM-To gain a better understanding on how the Fc⑀RI␤ ITAM functions in mast cell effector responses, we investigated its role in signal generation. We first analyzed the effect of wild type (YYY) and YFY, FYF, and FFF mutant ITAMs on Fc⑀RI phosphorylation and interaction with Lyn kinase to understand whether there was an association between the interaction with Lyn and the observed effects on mast cell degranulation and cytokine production. This line of investigation might also allow us to differentiate the relative importance of the noncanonical and canonical tyrosines in receptor phosphorylation and Lyn interaction. Tyrosine phosphorylation of Fc⑀RI␤ was reproducibly but minimally reduced in YFY cells; however, a striking inhibition of its phosphorylation was apparent in FYF and FFF cells (Fig. 4A), demonstrating that the canonical tyrosines were most important to this event. This was mirrored by a decrease in the phosphorylation of the ␥ chains in the same mutants, albeit not to the same extent (Fig. 4A). Because the decreased phosphorylation of the ␤ and ␥ chains suggested a possible loss of interaction with Lyn, we investigated the ability of the mutant ITAMs to co-immunoprecipitate Lyn kinase. Whereas the YFY mutant reproducibly showed a slight reduction in co-immunoprecipitated Lyn kinase, Lyn was not co-immunoprecipitated by the FYF and FFF mutants (Fig. 4B), suggesting that one or both of the canonical ITAM tyrosines was required for this interaction. Lyn kinase is also critical to the activation of Syk and thus to the phosphorylation of the Syk substrate, LAT. When we analyzed the phosphorylation of Syk and LAT, a minimal reduction of Syk, but not LAT, phosphorylation was seen in YFY cells (Fig. 4, C  and D) where Lyn could still bind to the ␤ ITAM (Fig. 4B). In contrast, the inhibition of Syk and LAT phosphorylation (ranging from 50 to 74%) was observed in cells transduced with the FYF and FFF mutants, which also failed to effectively bind Lyn (Fig. 4B).
We next explored the possibility that the activity or localization of the SHP-2 and the lipid phosphatase SHIP might be affected by mutation of the Fc⑀RI␤ ITAM based on the previous in vitro observation of their interaction with this receptor subunit (31,32). Investigation of their phosphorylation status in the YFY, FYF, and FFF mutants revealed that SHP-2 phosphorylation was relatively unaffected by mutation of the noncanonical (YFY) tyrosine and only modestly (the observed trend was not significant) reduced by mutation of canonical (FYF) tyrosines. Only when both canonical and noncanonical tyrosines were mutated (FFF) was significant inhibition of SHP-2 phosphorylation observed (Fig. 4E). In contrast, the phosphorylation of SHIP-1 was significantly inhibited by the mutation of the noncanonical tyrosine and this defect was maintained in the FYF and FFF mutants (Fig. 4F). Thus, SHIP-1 phosphorylation, which was shown as important for SHIP localization to the membrane (33), is positively regulated by both the Fc⑀RI␤ noncanonical as well as the canonical tyrosines.
We further explored whether the mutation of the Fc⑀RI␤ canonical or noncanonical tyrosines showed selectivity in interactions with SHP-2 or SHIP-1. We used phospho-ITAM peptides to enhance these interactions and allow detection of modest changes in possible interactions. As shown in Fig. 4G, SHP-2 was found to interact with the Fc⑀RI␤ phospho-ITAM equivalent (see "Experimental Procedures") of YYY and YFY but not with FYF and FFF. SHIP-1 also interacted with YYY and YFY and was very weakly but consistently detected with FYF. No interaction was observed with FFF. Interestingly, in this series of experiments, we also found that the p85 regulatory domain of PI3K also interacted with the Fc⑀RI␤ YYY and YFY ITAMs and was also weakly but consistently found to interact with FYF but not with FFF ITAMs. This interaction is consistent with the previous report of SHIP-1-mediated recruitment of PI3K (p85) during Fc␥RIIb1-mediated inhibition of the B cell antigen receptor (34). As expected, Lyn interacted with YYY and YFY but not with FYF and FFF ITAMs, confirming the co-immunoprecipitation experiments shown in Fig.  4B. However, the modest differences in the amount of Lyn binding to YYY and YFY ITAMs seen in the co-immunoprecipi- tation experiments were not observed, suggesting that other influences on Lyn interaction with the Fc⑀RI␤ are likely to account for the modest loss of Lyn interaction in the co-immunoprecipitation experiments from YFY-expressing cells. Syk failed to interact with the Fc⑀RI␤ ITAM, confirming the specificity of the demonstrated interactions (Fig. 4G).

Phosphorylation of PLC␥1/2 and the Calcium Response Is Also Impaired in Cells Expressing FYF and FFF Mutant
ITAMs-We recently demonstrated that LAT is required for PLC␥ activation and calcium mobilization in mast cells (26). Mutation of the four distal tyrosines of LAT revealed that the PLC␥ binding site (Tyr-136) is most critical for the tyrosine phosphorylation of PLC␥1 and 2, calcium responses, and degranulation (27). Thus, we explored the impact of the decreased LAT phosphorylation observed in FYF and FFF ␤ chain ITAM mutants and the status of PLC␥1 and PLC␥2 phosphorylation and calcium responses. A minimal but not significant decrease in the phosphorylation of PLC␥1 and 2 was observed in the YFY mutant (Fig. 5A). In contrast, a marked reduction (up to 70%) of the phosphorylation of these proteins was observed in FYF and FFF mutant cells (Fig. 5A). As shown in Fig. 5B, the marked inhibition of PLC␥1 and 2 phosphorylation in the mutant FYF and FFF ITAM-expressing cells caused a significant delay in the initiation of the calcium response with the time required for half-maximal rise of calcium in response to Ag increasing by 4 -6-fold (Fig. 5B). In contrast, the YFY-expressing cells showed no significant delay in the calcium response, although an effect on the extent of the calcium rise was noted in several experiments (Fig. 5B). Collectively, these findings demonstrate the contributory importance of the canonical ␤ ITAM tyrosines in events leading to the calcium response. The findings also suggest an association between the loss of receptor-associated Lyn and delayed calcium responses, because the observed signaling phenotype and a delayed calcium response were previously reported in Lyn-deficient mast cells (21,35). Moreover, the findings are consistent with the previously described amplification function of Fc⑀RI␤ on the calcium response (9).

ERK, p38 MAPK, IKK␤, and IB Phosphorylation Is Enhanced in Mutants (YFY and FFF) with Increased Cytokine
Production and Secretion-Given that many of the phenotypic traits of ␤ ITAM mutation mirrored those of Lyn deficiency (21,35,36), we explored the effect of ITAM mutation on MAPK activation, because the activity of these kinases is intact or enhanced in Lyn-deficient mast cells. Strikingly, all of the ITAM mutants showed an intact activation response for ERK, JNK, and p38 MAPK, suggesting that the activation of MAPK was essentially independent of the Fc⑀RI ␤ ITAM (Fig. 6, A-C). Moreover, a sustained and/or enhanced phosphorylation of ERK and p38 MAPK were reproducibly observed in YFY and FFF mutant cells when compared with wild type ITAM cells (Fig. 6, A and C). The phosphorylation of JNK and of protein kinase B/Akt, which is important for cell survival and gene expression, was also intact (Fig. 6, B and D). Thus, the path- ways that lead to gene expression were intact or enhanced (ERK and p38 MAPK) in Fc⑀RI␤ ITAM mutants.
IL-6 and TNF␣ mRNA production was demonstrated to depend on NFB activity in mast cells (37,38). Moreover, Krystal and colleagues (39) have recently shown that SHIP-1 Ϫ/Ϫ mast cells had increased NFB activity. Given that both receptorassociated Lyn phosphorylation and SHIP-1 phosphorylation are lost in the YFY and FFF mutants, we explored the phosphorylation/degradation of the components of NFB signaling. The binding of IB to NFB sequesters NFB in the cell cytoplasm. The phosphorylation of IB by a macromolecular kinase complex, comprised of two catalytic subunits (IKK␣/␤) and one regulatory subunit (IKK␥/NEMO) with adapter function, causes the degradation of IB and releases NFB, allowing its nuclear translocation. IKK␤ has been shown as essential for both the phosphorylation of the NFB subunit p65(RelA) and IB, whereas IKK␣ phosphorylates p65(RelA) (40). As seen in Fig. 7, the phosphorylation of both IKK and IB was enhanced in YFY and FFF mutant cells. In particular, a significant enhancement of IKK␤ phosphorylation was observed in the YFY mutant and the phosphorylation of this subunit was further enhanced and more sustained in the FFF mutant (Fig.  7A). Whereas some induction of IKK␣ phosphorylation was observed, it was not as dramatic as that of IKK␤. Moreover, the phosphorylation and degradation of IB were consistent with increased IKK␤ activity (Fig. 7B). The extent of reduction in IB protein mirrored the enhanced phosphorylation of IKK␤ and IB in the YFY and FFF mutants, respectively, demonstrating the cooperation between the noncanonical and canonical tyrosine residues of the Fc⑀RI␤ ITAM in negative regulation of the NFB pathway leading to cytokine production.
Whereas it is recognized that initiation of IKK signaling by a wide variety of stimuli results in NFB activation (41), we investigated whether cells transduced with Fc⑀RI␤ (YYY, YFY, FYF, and FFF) showed levels of nuclear NFB binding activity that reflected their respective levels of activation of IKK␤ and cytokine production. As a control, we measured the nuclear NFB binding activity of the nonstimulated transduced cells of each genotype and found no significant difference (Fig. 7C), consistent with the requirement of Fc⑀RI stimulation for induction of IKK␤ and IB phosphorylation. Fc⑀RI stimulation resulted in increased nuclear NFB binding activity for all of the cells expressing the exogenous Fc⑀RI␤. YFY and FFF mutants showed respective increases in the levels of nuclear NFB binding activity relative to the wild type (YYY) transfectant, as seen for IKK␤ and IB phosphorylation. In contrast, whereas the nuclear NFB binding activity of the FYF mutant also increases upon Fc⑀RI stimulation, this activity was decreased relative to the wild type (YYY) and mirrored the IKK␤ and IB phosphorylation for this mutant (Fig. 7, A and B). The increased NFB binding activity was previously demonstrated as essential for mast cell TNF␣ and IL-6 production (37, 39), thus

FIG. 4-continued
Negative Regulation of Cytokine Production by Fc⑀RI␤ collectively establishing the importance of the Fc⑀RI␤ in regulating these cytokines. DISCUSSION The Fc⑀RI␤ ITAM has long been recognized as an important site of interaction for Lyn kinase (15,42,43), and this subunit was demonstrated to provide an amplifying function in Fc⑀RI expression, IgE-dependent signals, and cellular responses (9,11). By employing retroviral transduction of Fc⑀RI␤-null BMMC with Fc⑀RI␤ ITAM mutants, we revealed a previously unrecognized dichotomy in Fc⑀RI␤-driven cellular response. Fc⑀RI-mediated degranulation and leukotriene secretion were adversely affected by the mutation of the canonical tyrosines of the Fc⑀RI␤ ITAM, whereas cytokine production and secretion were enhanced by mutation of either the noncanonical tyrosine or both the noncanonical and canonical tyrosines. Our findings demonstrate that Fc⑀RI␤ ITAM tyrosine residues that control the lymphokine responses also control the extent of activation of MAPK family members and the NFB signaling pathway. Cooperativity of the noncanonical and canonical tyrosines in control of cytokine production was apparent. This was most clearly demonstrated by the failure of the FYF mutant to enhance lymphokine responses and NFB signaling. In contrast, the enhancing effects of the YFY mutant on these responses were obvious but modest compared with the effect of the ITAM-null mutant (FFF). Therefore, whereas the noncanonical tyrosine plays the major role in exerting negative reg-ulation on NFB activation and lymphokine production, it requires the canonical tyrosines to amplify this effect.
Many of the phenotypic traits observed by mutation of the Fc⑀RI␤ ITAM are recapitulated in Lyn and SHIP-1-deficient mast cells. An analysis of signals in Lyn-deficient mast cells demonstrated a pronounced loss of Syk, LAT, and PLC␥1 and 2 phosphorylation (21,35,36). This is similar to the observed defects in FYF and FFF Fc⑀RI␤ ITAM mutants, which did not associate with Lyn, arguing that these defects may largely be a consequence of the loss of Lyn-Fc⑀RI␤ interaction. Similarly, calcium responses were also delayed and the extent of the response was diminished as seen in Lyn-deficient mast cells (21,35). In contrast, Kawakami and colleagues (35) demonstrated that MAPK activation was intact (JNK) or enhanced (ERK2 and p38 MAPK) (35) in Lyn-deficient mast cells mirroring the Fc⑀RI␤ ITAM-null mutant. A similar phenotype was seen with SHIP-1 deficiency, a protein whose phosphorylation requires Lyn activity (44,45). Krystal and co-workers (46) show enhanced ERK2 activity in these mast cells. However, in contrast to the partial defect of the Fc⑀RI␤ ITAM mutants on degranulation and leukotriene secretion, Lyn and SHIP-1-deficient mast cells hyperdegranulate (21,46). This finding suggests that the hyperresponsive degranulation of Lyn-and SHIP-deficient mast cells is not a consequence of the loss of negative regulation by Lyn at the level of the receptor itself, because the binding of Lyn to the FFF-mutated ␤ chain was not FIG. 5. The canonical tyrosines of the Fc⑀RI␤ are required for phosphorylation of PLC␥1 and 2 and for normal calcium responses. A, IgE-sensitized cells expressing the indicated Fc⑀RI␤ ITAM were stimulated with Ag for 3 min and lysed for immunoprecipitation with Anti-PLC␥1 or 2. Tyrosine phosphorylation was detected by antibody to phosphotyrosine (Anti-pY), and subsequently immunoblots were reprobed for protein loading. Fold induction was determined from three individual experiments by densitometry and normalized to the respective protein. **, p Ͻ 0.01 relative to wild type ITAM (YYY). B, cells expressing the indicated Fc⑀RI␤ ITAM were incubated with IgE and dual-loaded with 16 M Fluo-4-AM and 16 M Fura Red. Cells were stimulated with Ag (30 ng/ml) and changes in dye fluorescence with time were determined by flow cytometry at 37°C. Once the calcium response to Ag was diminished, cells were restimulated with phorbol 12-myristate 13-acetate and the calcium ionophore A233187 to determine a maximal response. Calcium mobilization is reported as the ratio of Fluo-4 to Fura Red fluorescence intensity over time. Half-maximal rise is reported as the time Ϯ S.D. (in seconds) required to reach 50% peak response to Ag stimulation for wild type and mutant ITAMs from four individual experiments. One representative of four individual experiments is shown.
observed and Lyn is active in SHIP-1-deficient mast cells. The findings also imply that the negative influence of Lyn and SHIP-1 on degranulation is separate from the positive or amplifying influence of the Fc⑀RI␤ ITAM. This is consistent with the findings that Lyn and SHIP are important in controlling cellular homeostasis (23,44,(47)(48)(49) and with the finding that Lyn deficiency increases Fyn activity that is required for degranulation (23). The resulting postulate is that activation of Fc⑀RI alone elicits negative regulatory controls that are independent of Lyn-Fc⑀RI interactions (23) but require active Lyn and Fc⑀RI aggregation.
Tyrosine phosphorylation of inositol 5Ј-phospatase SHIP-1 does not appear to be important for its activity; however, it plays an important role in membrane targeting of this phosphatase (33), probably because of its ability to interact with membrane proteins (similar to Lyn) and become stabilized in the membrane (50). Because the substrate of SHIP activity (inositol 1,4,5-trisphosphate) is membrane-restricted, the targeting of SHIP to the membrane is critical for its function in controlling the levels of inositol 1,4,5-trisphosphate and thus influencing the activity of inositol 1,4,5-trisphosphate-binding proteins such as the PI3K-dependent kinase 1, PLC, Akt, and others. Interestingly, the mutation of the noncanonical tyrosine (YFY) caused a significant loss of SHIP-1 phosphorylation, although SHIP-1 interaction with Fc⑀RI␤ and Lyn recruitment to Fc⑀RI␤ was still detected. This finding suggests that the phosphorylation of SHIP-1 rather than recruitment to the Fc⑀RI␤ might play a more significant role in the suppression of cytokine production. Remarkably, SHIP-1 was demonstrated to negatively regulate Fc⑀RI induction of IL-6 production by inhibiting NFB activity, which drives IL-6 transcription by binding to the NFB locus of the IL-6 promoter (39). The absence of SHIP was found to enhance IB phosphorylation and degradation as well as increase NFB DNA binding activity, similar to our observations in the YFY and FFF mutants. Collectively, these findings implicate Fc⑀RI␤-associated or -regulated SHIP-1 in promoting this phenotype. Although we exhaustively explored co-immunoprecipitation of SHIP with the Fc⑀RI␤, we could not reproducibly establish this interaction. This is probably due to the small amount of SHIP associating with Fc⑀RI␤. Nonetheless, our in vitro studies with phospho-ITAM peptides showed a strong SHIP-1 interaction with the canonical tyrosines and a modest, but reproducible, interaction when only the noncanonical tyrosine was phosphorylated.
We cannot formally exclude a conformational alteration of Fc⑀RI␤ engendered by ITAM tyrosine mutation. However, several of our findings argue against this possibility. First, the differential role of the Fc⑀RI␤ ITAM in degranulation and leukotriene secretion versus NFB signaling and lymphokine production argues that the effect is not a broad consequence of conformational inactivation or activation. Second, selectivity in activation (ERK2 and NFB) and inactivation (Syk, LAT, and PLC) was observed, demonstrating that the mutations of Fc⑀RI showed selectivity in their effects on molecules that interact with the receptor (Fc⑀RI␥) as well as others (similar to LAT) that do not require receptor localization for their activity or function (26,51). Third, no effects on Fc⑀RI expression were observed regardless of which tyrosine residue was mutated in the ITAM in contrast to a previous study in which Fc⑀RI␤ deletion mutants with altered conformations showed defective or variable expression (52). Finally, the phenotype observed for Fc⑀RI␤ mutant mast cells showed striking similarity to Lyn-or SHIP-1-deficient mast cells (21,23,35,39) and reproduced the previously demonstrated SH2 domain interaction of Lyn with receptor and the described weak signaling phenotype of the human ␣␥ 2 Fc⑀RI (8,9,43).
It is not yet clear whether the cooperativity of the noncanonical and canonical tyrosines in negative regulation of cytokine production is entirely mediated through Lyn and SHIP-1. The possibility of another player is suggested by the FYF mutant, which did not bind Lyn (similar to FFF), showed reduced SHIP-1 phosphorylation (similar to YFY and FFF) and its reduced binding to receptor but yet failed to enhance cytokine production. Identification of other molecules that specifically bind the noncanonical tyrosine may reveal the key negative regulatory component(s). Regardless, our findings show that the Fc⑀RI␤ can act not only as an amplifier of signaling leading to degranulation and leukotriene secretion but also as a negative regulator of selected signals that are important in gene expression. This presents the intriguing possibility that Fc⑀RI␤ ITAM can shift the balance between the amplifying and inhibitory functions of this subunit, thus influencing the allergic response.